AMD CDNA architecture (CDNA-CDNA4)

AMD CDNA architecture (CDNA-CDNA4)#

ROCm Compute Profiler makes available an extensive list of metrics to better understand achieved application performance on AMD Instinct™ MI-series GPUs including CDNA™ architecture based GPUs like the AMD Instinct MI100, CDNA2 architecture based GPUs such as the AMD Instinct MI210, MI250, and MI250X, CDNA3 architecture based GPUs such as the AMD Instinct MI300A, MI300X, and MI325X, and CDNA4 architecture based GPUs such as MI350X and MI355X.

Note

  • For AMD Ryzen™ / RDNA™ APUs (e.g. gfx1151/RDNA3.5), see RDNA3.

  • For top-level metrics details on CDNA and RDNA architecture, see Performance model.

The table provides key details and support available for the different CDNA architectures:

✅: Supported ❌: Unsupported

Architecture details

Architecture

CDNA

CDNA 2

CDNA 3

CDNA 4

Chip packaging

Single Die

Up to two graphics Compute Dies (GCDs) into single package.

One logical processor with dozen chiplets, configurable with partition modes.

Similar to CDNA3, Multi-Die chiplet, but with two I/O Dies (IODs)

Supported series

MI100

MI210

MI300A

MI350X

MI250

MI300X

MI355X

MI250X

MI325X

Spatial partition mode

❌

❌

Compute partition mode and Memory partition mode

Compute partition mode and Memory partition mode

Data type support

Architecture

FP32

FP64

FP16

INT32 ADD/LOGIC/MAD

INT8 DOT

INT4 DOT

FP32 GEMM

FP64 GEMM

FP16 GEMM

BF16 GEMM

INT8 GEMM

Packed FP32

TF32 GEMM

FP8/BF8 GEMM

CDNA

✅

✅

✅

✅

✅

✅

✅

❌

❌

❌

❌

❌

❌

❌

CDNA2

✅

✅

✅

✅

✅

✅

✅

✅

✅

✅

✅

✅

❌

❌

CDNA3

✅

✅

✅

✅

✅

✅

✅

✅

✅

✅

✅

✅

✅

✅

CDNA4

✅

✅

✅

✅

✅

✅

✅

✅

✅

✅

✅

✅

❌

✅

To best use profiling data, it’s important to understand the role of various hardware blocks of AMD Instinct GPUs. Refer to the following top-level GPU architecture diagram to understand the hardware blocks of each architecture.

CDNA top level architecture diagram with zoomed view of Compute unit
CDNA2 top level architecture diagram with zoomed view of Compute unit
CDNA3 top level architecture diagram with zoomed view of Accelerator Complex Dies (XCDs)
CDNA4 top level architecture diagram

This section describes each hardware block on the GPUs as interacted with by a software developer to give a deeper understanding of the metrics reported by profiling data. Refer to Profiling by example for more practical examples and details on how to use ROCm Compute Profiler to optimize your code.

Note

In this documentation, MI2XX refers to any of the CDNA2 architecture-based MI200 series GPUs, such as AMD Instinct MI250X, MI250, and MI210 GPUs interchangeably in cases where the exact product at hand is not relevant. For product details, see AMD Instinct GPUs.

For a comparison of AMD Instinct GPU specifications, refer to Hardware specifications.

Hardware block chapters#

The AMD Instinct performance model is divided into the following blocks:

Contents